Tumor Cell Gene Correction Bacterium Conjugation It is estimated that 13 percent of colorectal cancers in the Western world are associated with defects in DNA mismatch repair pathways. About half of these originate as inherited defects, such as in hereditary non-polyposis colorectal cancer (HNPCC). Studies of HNPCC families have led to the association of tumors with mutations in any of five DNA mismatch repair genes. One long-range goal of the proposed work would be to prevent the onset of colorectal cancers such as HNPCC by early in vivo correction of genetic defects by gene replacement therapy, using normal flora bacteria as gene donors. I have published experiments establishing bacterium-to-mammalian cell conjugation, and propose here to apply this novel method of gene delivery to correct tissue culture cells defective in the gene most commonly associated with HNPCC. Experiments will first determine the targeting efficiencies of conjugative gene delivery, as compared to transfection, using a system developed to correct CHO cells defective in two different indicator genes. Correcting one gene will indicate random insertion; correcting the other will require targeting by homologous recombination. Conjugation, which supplies single-stranded DNA, is expected to out-perform transfection in targeting by homologous recombination. Next, bacterial conjugation will be used to correct mouse embryonic fibroblasts defective in mMIhl, the murine homologue of the human gene (hMLH 1) commonly associated with HNPCC. Lastly, conjugation will be used to deliver hMLH1 to human colon tumor cells (HCT116). Delivery of genomic and cDNA forms of the gene will be compared, and it is expected that the genomic form will be necessary for proper gene expression and restored phenotype. These experiments will provide good preparation for the development of bacterial conjugative transfer designed to correct defective genes in vivo.